ABSTRACT
The work presented in this chapter arose from two questions that may at first appear unrelated: should we accept the premise that skilled word reading is supported by lexical and sublexical routes of processing, and how can we model the learning of orthographic and phonological representations for multisyllabic words? The field as a whole has answered “yes” to the first question, but is mostly silent with regard to the second. Here I will consider an alternative to lexical and sublexical routes, one in which these are modes of processing instead of components of processing. My colleagues and I have pursued this idea by exploring theories of lexical processing in which orthographic, phonological, and semantic codes are all mediated by one level of representation. In our pursuit, we were forced to confront the long-standing problem of modelling multisyllabic words. This chapter is an account of our efforts that have recently culminated in the junction model of lexical processing.
